On Mon, 21 Jan 2002, mikel_ind wrote:
> --- In APBR_analysis@y..., "thedawgsareout" <kpelton08@h...> wrote:
> > .....points per shot defined rather simply as points/field goals
> > attempted.
> > .....points from field goal attempts divided by shots,
> > or (points-free throws made)/field goals attempted.
> > Which is the generally accepted definition for this stat? Has it
> > changed over the years?
> Not sure which of these measures predates the other. Counting points
> from FT and dividing by FGA sounds bizarre to me, as though FT were
> just bonus points (and-1). I haven't seen anyone passing this off as
> a useful stat.
Yeah, it was bizarre to see that stat surface (I don't remember seeing it
prior to a year or two ago, but maybe it's been around for awhile). A
really dumb statistic. If it becomes popular, it will set basketball
stats back a couple of decades. It's as lame as the "RBI + RUNS - HR"
statistic for "runs created"; not that "RBI+RUNS" is so bad, it's just
that it's so stupid to subtract the home runs. Similarly, as you say,
adding FTM and dividing by FGA is bizarre.
> Subtracting FT to get "points from FG" and dividing by FGA, gets you
> the "effective FG%", or whatever you wish to call it, which is an
> improvement on the old FG% that still appears in all standard
> lists. A player who shoots at least 2/3 as well from 3-range as he
> does from 2-range improves his effective % by shooting more 3s.
> I have some gigabytes of stats that assume 2 FTA = 1 FGA, for
> purposes of calculating a "combined shooting pct", "scoring
> pct", "scoring efficiency", or whatever we end up calling it:
> Pct = Pts / (FGA*2 + FTA)
Yes, this is the much better statistic to use. I like to call it by the
third name you list: "scoring efficiency". I believe there used to be
a website, now I think defunct, where a guy calculated and claimed to have
copyrighted or trademarked the term. If so, an idea both greedy and dumb
(dumb to trademark the term that it, not dumb to conceive of the
statistic). I like to abbreviate this stat as "PT%".
> This couldn't properly be called "points per shot", but
> rather "baskets per shot", roughly speaking. In other words,
> pts/shot should be around 1.000, and 2pts/shot should be around .500.
> Dean Oliver has used a conversion rate to multiply by the FTA in the
> above formula. Because of the and-1 FT and technicals, not all FT
> are a separate scoring attempt. This makes good sense.
Yes, although very tough to do without play-by-play descriptions of
individual plays. Does Dean O convert individual FTs, or use an overall
universal conversion factor for all players' FTs? I once estimated that
the conversion factor for FTAs should be about 12/13 or 13/14 in the NBA,
but that was a very rough estimate.
> Applying this formula to eras past, one might consider the days of
> the 1-shot foul, the 3-to-make-2, and so on, and use a varying FT
> conversion factor.
Yes. Boy that'd get really complicated.